CB-ICA: a crossover-based imperialist competitive algorithm for large-scale problems and engineering design optimization

Optimization is increasingly important due to its application in the real-world problems. The imperialist competitive algorithm (ICA) is a successful optimization algorithm in many applications. However, in solving complex problems, especially in high dimensions, ICA easily falls into local optimal and experiences the premature convergence. In this work, a crossover-based imperialist competitive algorithm (CB-ICA) was proposed for solving this problem. The proposed algorithm faces three changes compared to ICA. To increase the exploration ability, the uniform distribution crossover and levy mutation methods were used in assimilation and revolution steps, respectively. Furthermore, the use of uniform crossover in the imperialist improvement step with appropriate data exchange leads to the improvement of convergence speed toward an optimal solution. The results of trials on 10 unconstrained large-scale tests show the advantage of CB-ICA over ICA and also over the related methods in terms of quality of results, convergence speed and reliability. In addition, the proposed method can be used in solving real-world problems with the use of constraint-handling technique. For this reason, CB-ICA has been compared in five engineering design problems with several state-of-the-art algorithms and obtained acceptable results.